Textural or Textual: How Vision-Language Models Read Text in Images | Read Paper on Bytez